I'm looking for some references regarding integration of DAEs in the form

$M(t) \frac{dy}{dt} + G(y(t),t) + f(t) = 0, \quad y \in R^n, M(t) \in R^{nxn}$

using a high order implicit Runge Kutta scheme. I'd like also to find some discussions about the stability of a RK scheme in its integration.

For the curious these DAEs derive from the application of the continuous adjoint method.